File size: 1,217 Bytes
bbde80b |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 |
# 3D Human Pose Estimation
## Data
1. Download the finetuned Stacked Hourglass detections and our preprocessed H3.6M data (.pkl) [here](https://1drv.ms/u/s!AvAdh0LSjEOlgSMvoapR8XVTGcVj) and put it to `data/motion3d`.
> Note that the preprocessed data is only intended for reproducing our results more easily. If you want to use the dataset, please register to the [Human3.6m website](http://vision.imar.ro/human3.6m/) and download the dataset in its original format. Please refer to [LCN](https://github.com/CHUNYUWANG/lcn-pose#data) for how we prepare the H3.6M data.
2. Slice the motion clips (len=243, stride=81)
```bash
python tools/convert_h36m.py
```
## Running
**Train from scratch:**
```bash
python train.py \
--config configs/pose3d/MB_train_h36m.yaml \
--checkpoint checkpoint/pose3d/MB_train_h36m
```
**Finetune from pretrained MotionBERT:**
```bash
python train.py \
--config configs/pose3d/MB_ft_h36m.yaml \
--pretrained checkpoint/pretrain/MB_release \
--checkpoint checkpoint/pose3d/FT_MB_release_MB_ft_h36m
```
**Evaluate:**
```bash
python train.py \
--config configs/pose3d/MB_train_h36m.yaml \
--evaluate checkpoint/pose3d/MB_train_h36m/best_epoch.bin
```
|